July 5, 1958
KINETICSOF Ortho-CLAISEN REARRANGEMENT
Strong interaction of this nature is possible since the carbons of the allyl group may lie in a plane directly over the aryloxy grouping so t h a t the acarbon of the allyl group is over the oxygen of the aryloxy moiety, the P-carbon is over the 1-position of the ring, and the y-carbon over the 2-position. This places one terminal carbon of the allyl group over the oxygen from which it breaks away and the other over the o-carbon to which it becomes bonded. This complex must be very short-lived if it is to maintain its orientation. The substituent data are explained if structure I1 contributes more to the resonance hybrid than 111-a positive charge which is stabilized by electron donation is placed on the aromatic ring during reaction and as a final stage, if equation 2 is to apply, a negative charge is cancelled a t the position meta to the substituent requiring electron withdrawal. T h a t I1 actually would contribute more than I11 is indicated by the fact that the difference between the ionization potential of the benzyl radical32and the electron affinity of the allyl radical33is about 17 kcal./mole greater than the difference between the ionization potential32of the allyl radical and the electron affinity of the benzyl radical.33 Although this comparison is not precise because the actual system involves a phenoxy radical rather than a benzyl radical, it is likely that the comparison is justified since the electron affinity of phenoxy radical34 is even less positive than t h a t of benzyl radical by 14 kcal./mole. Finally, this latter mechanism is supported by an analogy to the Diels-Alder reaction for which a charge transfer complex mechanism has been postulated by W0odward.~5 Like the (32) D. P. Stevenson quoted by A. Streitwieser, Jr., Chem. Reus., 66, 571 (1956). (33) N. S. Hush and K. B. Oldham quoted by H. 0. Pritchard, ibid., 63,529 (1953). (34) N. S . Hush quoted in H. 0. Pritchard, ibid., 62, 529 (1953). (35) R. B. Woodward, THIS JOURNAL,62, 3058 (1942); R. B. Woodward and H. Baer, ibid., 66, 645 (1944).
[CONTRIBUTION FROM THE
3277
Claisen rearrangement, the effect of substituents on the Diels-Alder reaction of l-(p-X-phenyl)-1,3butadienes and maleic anhydride has been found to be most suitably correlated by use of the u + constants and a small, negative p.23.36 The existence of a polar intermediate or transition state in the Claisen rearrangement is demonstrated not only by the substituent effects but also by the preliminary results on the effect of solvent (Table VI). A semi-quantitative treatment of the data is not possible since the dielectric constants of these solvents have not been measured a t the temperature of rearrangement, and the change of dielectric constant with temperature is different for different solvents. However, it is encouraging that the solvent of lowest dielectric constant a t ordinary temperatures (n-octane) yields a rate onethird to one-seventh that obtained with the other solvents, and that the addition of a salt to Carbitol (which should increase the dielectric constant) increases the rate. The solvent rate effect observed for the Claisen rearrangement is similar in magnitude to t h a t for the Menschutkin reactions of methyl iodide in benzene-nitrobenzene mixtures, 37 a reaction which is most definitely polar in nature. There are, therefore, a t least three reasonable mechanistic explanations for the substituent and solvent effects observed in the Claisen rearrangement. Unfortunately, a t present, there is not enough experimental information available to decide among these possibilities. (313) The effect of a y-methyl group in the allyl portion of the ether (footnote 30) can be explained as being due to a larger contribution by structure I11 to the hybrid. The additional methyl group should tend to stabilize 111. (37) H. C. Raine and C. iT.Hinshelwood, J . Chem. SOL, 1378 (1939); K. J. Laldler and C. N. Hinshelwood, ibid., 858 (1938).
COLUMBUS 10, OHIO
DEPARTMENT OF CHEMISTRY
O F THE UNIVERSITY O F
I IS CONS IS]
A Kinetic Study of the ortho-Claisen Rearrangement' BY HARLAN L. GOERINGAND ROBERTR. JACOBSON RECEIVED JANUARY 27, 1958 The relative rates of the ortho-Claisen rearrangement of allyl phenyl ether, fifteen m- and p-substituted allyl phenyl ethers and a-,0- and y-methyallyl phenyl ether at 185 and 197" in diphenyl ether have been determined. A dilatometric method for following the reaction accurately using low concentrations (about 0.1 M ) of substrate was developed. The relative rates of rearrangement of allyl p-cresyl ether in twelve solvents also have been determined. The relative reactivities of the series of p-substituted allyl phenyl ethers can be correlated by the CT+ substituent constants.
Introduction
The relative positions of the atoms in the reactants and products are the same for the two reactions In many respects the ortho-Claisen rearrange- and in each case the over-all reaction can be summent2 is similar to the intramolecular (SNi') re- marized as a six-membered cyclic process as illusarrangement of allylic ester^.^ Both of these reac- trated by I and II.4 In the Claisen rearrangement tions are first-order intramolecular p r o c e s s e ~ . ~ , ~ (1) This work was supported in part by the Office of Ordnance Rcsearch, U. S. Army, and in part by thP Research Committee of the
Graduate School with funds given by the Wisconsin Alumni Research Foundation. (2) D. S. Tarbell in R . Adams, "Organic Reactions," Vol. 11, John Wiley and Sons, Inc., New York. N. Y., 1944, Chapt. 1. (3) (a) H. L. Goering and R. W. Greiner. THIS JOURNAL, 79, 3464
(1957); (b) H. L. Goering, J. P. Blanchard and E. F. Silversmith, ibid.. 76, 540b (1954); H. I.. Goering and E. F. Silversmith, ibid., '77, 1129 (1955); (c) unpublished work of R. H. Jagow and M. M. Pombo. (4) It recently has been shown with 0'5-labeled esters that in the SNi' rearrangement of allylic esters the carbonyl oxygen atom in the reactant becomes the alkyl oxygen atom in the product as illustrated by 11; W. E. Doering, private communication (1955); E. A. Braude and D. W. Turner, Chemirlvy ?-f Industry, 1223 (1955).
HARLAN L. GOERING 4 N D ROBERT R. JACOBSON
3278
(I) the initially formed dienone is converted to the corresponding phenol.
Vol. 80
there is evidence that there is some covalent bonding between the allylic carbon atoms and the oxygen atoms in VI. It has been suggested* that similar intermediates may be involved in other
I
K \,- I
I1
It would appear that the stereochemical relationship between reactant and product should be the same for the two reactions, however, there is evidence that this may not be the case. It has been d e ~ n o n s t r a t e d ~that ~ J ~ in the rearrangement of trans-a, y-dimethylallyl esters (HI), the configuration of the asymmetric carbon in the product is opposite that of the asymmetric carbon in the reactant, ;.e., in this system the product and reactant are enantiomers as illustrated below. On the other hand, it has been conclude@ t h a t in the H
H’
13
‘11
rIIa
/
11
li
IIlb
Claisen rearrangement of optically active trunsa ,y-dimethylallyl phenyl ether (IV),6 the configurations of the asymmetric centers in the reactant and CHaCHCHzCHCHa I
product (V) are the same. If this conclusion is correct, the stereochemical relationship between reactant and product differs for the two cases. I n another problem the relative configurations of IV and V are being re-examined t o determine whether this apparent discrepancy is real. Considerable information concerning the timing and nature of the electronic shifts in the SNi‘ rearrangement summarized by I1 has been presented in earlier papers3 It appears t h a t an intermediate V I (an “internal ion-pair” intermediate according to Winstein’s classification’) is involved in this reaction as follows. As indicated previouslyaa ( 5 ) H. Hart, THISJ O U R N A L , 76, 4038 (1954). (6) E. R . Alexander and R. W. Kluiber, ibid., 1 3 , 4304 (1931). (7) S. Winstein, E. Clippinger, A. Fainberg, R. Heck and G. Robin son, ibid., 78, 328 (1956).
12
intramolecular processes including the Claisen rearrangement. I n order t o obtain inforniation concerning mechanistic details of the ortho-Claisen rearrangement (I) we have examined the effects of substituents and variation of solvent on the rate of rearrangement. The results of this investigation are presented in this paper. The rates of the ortho-Claisen rearrangement of a few aryl allyl ethers in the absence of solvent have been reported previously.”l” Kinetic studies of the rearrangement of allyl /]-cresyl ether in diphenyl ether lo and allyl 2,,l-diinethylphenyl ether in diethylaniline“ also have been described. I n the latter two cases rather high concentrations of substrate were used (5-335; by weight of the reaction mixture) and the reactions were followed by determination of the amount of phenol formed by quantitative acetylation. This method is liniited to non-hydroxylic solvents and probably could not be used with low concentrations of substrate. I n the present work a dilatometric iiiethotl was developed which can be used to follow reactions a t temperatures up to 200”. Previously dilatoiiietry has been used to follow reactions only a t ternperatures near room Tletails of this method are included in the Experimental section. \Vith it the ortho-Claisen r e a r r q + n c n t can be followed with a high degree of accuracy a t teniperatures of iS3-800” using low concentratioiis of substrate (about 0.1 111). l y e have used this niethotl to deterniine (a) the relati1.e reactivities of allyl phenyl ether and eighteen substituted allyl phenyl ethers and (b) the relative rntes of rearrangenicnt in twelve solvents. Experimental Allyl Phenyl Ethers.-Except as iiotecl bel(^^ a l l o f tlic ethers were prepared bj- the following tnethotl which is u modification of the method used by Claiseii.I3 A solutioti of 0.2 mole of plieiiol, 0.22 inole of ally1 brotnitlc atid 0.22 mole o f potassium carbonate in 50 1111. of drl- acetoiie W:LS refluxed for 8 hours. hfter cooling, the reictimi inisture was diluted with 250 nil. of water aiid extracted with three 35-ml. portions of ether. The organic layer WIS ivashetl first with 105; aqueous sodium hydroside aiid then with water and dried over magricsiurn sulfate. .lfter reinoviil o f the solveiit the residue w a s fractioiiatcd under rcdueetl ( 8 ) I) J . Cram, t b i d , 7 4 , 2129 (1952); 75, 332 (10;iX). (9) €1. Schrnid and K , S c h m i d , H ~ l a .Chzin. .4c/ii, 35, 18i!l (1932). (10) J . F. Kincaid a n d D. S. Tarbell, 7‘rirs J O I J R N A I . , 61, 30% (1039). (11) 1’. Kall,erer and H. Schrnid, 11~1:’. i ’ h i i i i . t l c l u , 40, 13 (1‘337) (12) (a) J . S . Brbnsted, el ( I L , , THISJULIRNAI., 49, 2584 (liI27); 6 1 , 428 (l!J?!I); (h) L. K . J . T o n g and A . 12.Olson, ibid., 66, 1701 (19.43). (13) L. Claisen and 0. Eisleh, .41n>~, 401, 21 (1913).
KINETICS O F O?'thO-CLAISEN REARRANGEMENT
July 5 , 1958
3279
TABLE I PHYSICAL PROPERTIES OF SUBSTITUTED ALLYLPHENYL ETHERS Obs. b.p.
Substituent
OC.
bIm.
Obs.
Lit. b.p. OC. Mm.
?$~SD
Lit. n b
Ref.
16 97.2-97.4 1 1.5679 136 11 0 119 11 M.p. 44.2-45.0" ..... .. M.P. 39-40' 35 70.0-70,5 0.1 1.5249 119-120 13 9, 10 100.0-100.5 9 1.5157 65 4 1,5199 (200) 13 106-107 12 97.0-97.3 10 1.5348 p-c1 13 110.0-110.1 10 1.5583 126 14 p-Br 73 11 1.5190 (26") 16, 67.5-67.6 10 1.5196 None 13 100.0-100.1 156 10 0.6 1.5258 P-COOCsH, 13 10 2 1.5069 142 106.0-106.5 p-CHO 89-90 0.6 1.5525 146-147 10 1.5432 (45") p-COCH3 >; 94.0-91.5 .3 M.p. 43.6-44.1" ..... .. ALP. 43" P-CN 13 105-106 .4 1.5789 160 12 P-NOz 35 125-126 15 66.0-66.1 .5 1.5263 WZ-OCH~ 136-137 8 M.p. 3 1 , 5 3 2 ' 92.2-92.3 .6 M.p. 32-33' m-SOz i 75.2-75.8 10 1.5136 70 8 1.6168 (20') p-CH3 9 87.1-87.5 8 1.5183 95-98 12 1.5187 (19') -(-CHJ 72 0.8 1.5110 (20') 59.8-60.0 0.4 1.5060 a ,-,-Dimethyl 6 a S.M. McElvain and E. L. Engclhardt, THISJOURNAL, 66, 1077 (1944). C . D. Hurd and L. Schmerling, ibid., 59, 107 M. \ir. Partridge, J . C h e w Soc., 3043 (1949). Mi. C. (1937). R. T. ilrnold and J. C. McCool, ;bid., 64, 1315 (1942). 9.R. Bartz, R. F. Miller and R. ;Idams, ibid., 57, 371 (1935). Wilson and R. Xdams, THISJOURSA-., 45, 528 (1928). Q J. v. Braun and W. Schirmacher, H e r . , 56, 538 (1923). P-NH2 P-OCzH, p-OCH3 P-CHB
PHYSICAL PROPERTIES ASD Substituted phenyl allyl ether
p-N(CH3)z P-CZHL WZ-COOC~H~ a-CH3
.~NALYTICAL
TABLE I1 DATAFOR ARYLALLYLETHERS AND SUBSTITUTED 2 ALLYLPHENOLS
B.p.
oc.
him.
ii25D
Empirical formula
Carbon, % Calcd. Found
Hydrogen, % Calcd. Found
88.1-88.2 96.0-9G. 8 88.0-88.1 42 9-43.0
0.5 9 0.2 0.8
1.5500 1,5124 1.5169 1.5078
CiiHlbNO CiiH140 C12H1103 CioH120
74.55 81.45 69.89 81.04
74.89 81.26 69.80 80.70
8.53 8.70 6.84 8.16
8.24 8.78 6.71 8.10
85 63.0-63 5 M.P. 83.0-84.0 M . p . 151-152
0.1 0.2
1.5681 1.5301
CiiH1SNO CilHliO CioHsKO CmH14Op
74.55 81.45 75.39 69.89
74.67 81.43 75.55 69.95
8.53 8.70 5.69 6.84
8.60 8.58 5.48 7.00
Substituted 2-allylphenols
4-K(CH3)2 4-CzHj 4-CN 5(or ~ ) - C O O C Z H ~
pressure with a short Vigreux column. The main fraction T h a t the samules were homoneneous was indicated bv the fact that fractionation g a v e fractions having the -same was refractionated prior to use in the kinetic experiments. physical properties. A11 of the samples had the correct T h e phenols were purified by distillation or recrystallization. nz-Carbethoxyphenol and p-carbethoxyphenol were chemical composition (carbon-hydrogen analysis) and the prepared by esterification of the corresponding liydroxy- infrared spectra showed the samples to be free of phenol. T h e structures of these several substituted allyl phenyl benzoic acids. p-Ethoxyphenol was prepared from hydroquinone and diethyl sulfate by the method used to prepare ethers reported in Table I were confirmed by oxidationi8 (as its anisole from phenol and dimethyl ~ u 1 f a t e . l ~@-Cyano- to the corresponding aryloxyacetic acids-p-NHz acetyl derivative), p-OC2Hj, p-OCHa, p-CI, p-Br, p-ii02, phenol was prepared from p-aminoplienol by the method nz-NO:! and -pCH.,. a,y-Dimethylallyl phenyl used previously to convert p-toluidine to p - t u l ~ o n i t r i l e . ~ ~WL-OCH,~, ether (presuinably the trans isomer6)), was oxidized to aT h e rest of the plietiols are available commercially. Allyl . (lit.19m.p. 115-116') p-aniinophenyl ether was prepared by hydrolysis of allyl phenoxyprupionic acid, m . ~ 115-116O in good yield. T h e structures of the p-COsC?H,, p-CHO p-acetaniitioplieiiy-1 ether.lG and p-CN phenyl allyl ethers were verified by conversion to Allylic chlorides were used for preparing CY- and -,-methylallyl and a,-,-dimethylallyl phenyl ether. These chlorides p-allyloxybenzoic acid.13 had the following physical properties: -pmeth>-lallyl chloThe new allyl phenyl ethers are shown in the upper half of ride, b.p. 83-83.7", %%D 1.4325 (lit.I7b.p. 83', n Z o1.4351); ~ Table 11. The structure of a-methylallyl phenyl ether was a-methylallyl chloride, b.p. 62.5-63.5', n l j ~1.4125 (lit.17 established by reduction (palladium-on-charcoal) to s a b.p. 63.5", 1.4150); a,y-dimethylallyl chloride, b.p. 27" butyl phenyl ether. The physical properties and infrared (30 mm.), TZ%D 1.4306 (lit." b.p. 18-20" (12-13 nim.), spectrum of this product were the same as for a n authentic n% 1.4311). Commercialll- available 0-methylallyl chlo- sample.20 The unsaturated ether also was oxidized to aride was used to prepare P-methylallyl phenyl ether. Ethanol phenoxypropionic acid, m.p. l15-l16°.19 1Vlien heated was used as the solvent instead of acetone for preparing the ether rearranged to o-crotylphenol (isolated in 947, allyl p-formylphenyl ether.13 yield), b.p. 67-68' (0.6 mm.), phenylurethan m.p. 66-67" Most of the allyl phenyl ethers used in the present work (lit.21phenylurethan m.p. 65-66"). have been described previously. These are given in Table I The structure of allyl p-ethylphenyl ether was established together with the observed and reported physical properties. (14) G. S . Hiers and F. I).Hagcr, "Organic Syntheses," Coll. Vol. I , 2nd edition, John Wiley and Sons, Inc., New York, N. Y., 1941, p. 58. (15) H . T. Clarke and R . R . Read, ibid., 11. 514. (16) L . Claisen, A > t n . , 418, 69 (1'31'3). (17) E. H . Huntress and E. E. Toops, J r . , "Organic Chlorine C o n pounds," John Wiley and Sons, Inc., S e w Y o r k . N.y . , 1948.
(18) W. M . Lauer and W. F. Filbert, THISJ O U R X A L , 68, 1388 (1936). (19) C. A. Bischoff, B e y . , 33, 925 (1900). (20) M. M. Sprung and E. S . Wallis, Tiirs J O C J R N A I . , 56, 1715 (1934). (21) I,. Claisen and E. Tietze, Be?., 59, 23-14 (1926).
3280
HARLAN L. GOERING AND ROBERT 12. JACOBSOS
by oxidation t o fi-ethylpheiioxyacetic acid, 1ii.p. 96.396.5' (lit.22 m.p. 97"). The structure of m-carbethoxyphenyl allyl ether was established by hydrolysis t o nz-allyloxybenzoic acid, m.p. 80.0-80.5'. 23 A n a l . Calcd. for CIDHIOO~: C, 67.40; H, 5.66. Found: C, 67.62; H, 5.59. The structure of the latter compound mas established by oxidation to nz-carboxyphenoxyacetic acid, m.p. 207-208" (lit.24m.p. 206-207"). Allyl p-dimethylaminophen) 1 ether was prepared in 33 % yield by alkylation of the p-amino ether by a method that has been used previously for alkq lating aromatic amines.*j The structure of this compound was established by conversion to p-dimethylaminophenol, m,.p. 77.5-78.0" (lit.26 m.p. 78") by refluxing 1 g. of the dimethylamino ether in 7 tnl. of 48'/c hydrobromic acid for 5 hours. T h e amino phenol nas isolated in 757* yield. Rearrangement of Allyl Phenyl Ethers.-Four of the allyl phenyl ethers listed in Tables I and I1 rearrange to phenols that have not been described in the literature. These new phenols are included in Table 11. The phenols were obtained from the ethers by the procedure given below for the conversion of allyl nt-carbetlioxyphenJ.1 ether to rearrangement product. This is one of the t n o cases where two rearrangement products are possible and the only case
F
1 Fig. 1.-Dilatometer
L
used in kinetic experinients.
(22) S. RI. RlcElvain, "The Characterization of Organic Compounds," revised edition, The hlacmillan Co., R-ew York, S . Y . , 1953, p. 268. (23) Evidently the melting point of 1 1 8 O reported for this compound by Scichilone, Gazz. c h i m . dd.,1 2 , 453 (1882), is incorrect. (24) R. Meyer and C. D. Duxzmal, Ber., 46, 3360 (1913). ( 2 6 ) W, L. Borkowski and E. C. Wagner, J. Oug. Chem., 17, 1128 (1932). ( 2 6 ) H. Wieland, Ber., 43, 712 (1910).
VOl.
so
in which t w o products are formed. .i solution of 3.0 g. of allyl m-carbethoxyphenyl ether in 6 g. of diphenyl ether was heated under nitrogen for 9 hours at 220'. After cooling, the solution was extracted with lOy0 sodium hydroxide and the aqueous layer was saturated with carbon dioxide and extracted with ether. Removal of the ether gave a residue, 2.95 g. (98%), which apparently was a mixture of the two possible rearrangement products, i . e . , 2-allyl-5-carbethoxyphenol and 2-allyl-3-carbetlioxyphcnol. The higher melting component of the mixture was purified by several recrystallizations. This material is included in Table 11. The lower melting component was not purified. Kinetic Exper;ments.-The dilatometer used to follow the reactions at temperatures of 185 arid 197' is illustrated by Fig. 1. This dilatometer, which was designed for use a t high temperatures, differs substantially from thosc used previously for following reactions a t much lower teinpcratures (near room temperature).12,27 The important features of this design are: (a) T h e coiled tube provides a high surface to volume ratio and thus temperature equilibration is rapid. This coil was constructed of 7 mm. thin walled Pyrex tubiug and had a capacity of about 30 nil. (b) The level of the reaction mixture (sealed in the dilatometer at or near room temperature) in the capillary (D) can be adjusted so as to be on scale at the reaction temperature. T h e capillary (D) was 10 inches long and had a precision bore 0.0224 inch in diameter. The length of the dilatometer from the bottom of the coil to the top of the capillary was about 14 inches. During the kinetic experiments the sealed portion of the apparatus (from A to C) was immersed in the constant temperature bath and held in place by clamps attached to the two glass rods (E and F) which extended through its cover. The constant temperature bath consisted of a Pyrex jar in a well insulated box. T h e box had a double paned window for observing the scale on the dilatometer. Except when in use the window was covered by a well insulated door. The Pyrex jar (12" in diameter and 18" deep) contained 6 gallons of Dow KO. 550 silicone oil. The bath was heated by two 500 watt heaters. One of t1ie.e served as a constant heater and the other as the intermittent heater. The lattcr was connected in series with a 50-watt light bulb. The constant heater mas connected to a variable viiltage control and the current for this heater was controlled by L: Raytheon line voltage regulator. The temperature of the bath was controlled by a carefully selected Philadelphia type S\Y 912 thermoregulator. The bath was also provided with an eficient stirrer arid a tight-fitting cover. The temperature fluctuations of the bath were less than 0.003" during a rate run and during the period of its use the absolute ternperature varied less than 0.05'. I n the kinetic experiments the dilatometer was fillcd as follows. Both A and C were sealed and the reaction mixture n-as added through B at room temperature. After filling to the top of the coil, B was sealed aud a scale with 2 mm. divisions was attached to the capillary tube. For emptying and cleaning the dilatometer A, B and C were opened. The level of tlic liquid in the capillary a t the reaction temperature was adjusted as follows. The sealed tiilatomcter was placed in the bath which had previously been heated a few tenths of a degree abovc the operating ternperature. After the excess liquid had overflowed into the reservoir (G) the temperature was lowered to normal and the level of the liquid dropped to the top end of the scale. This operation required about 5 minutes. After equilibration for 10 minutes a t the reaction temperature the zero-time reading (Ro) was taken and the rate of decrease in volume was determined. The total drop during a kinetic cxperitnent varied from 50 to 90 mm. and the readings were recorded to 0.2 mm. The readings were always taken a t the same time during the heating cycle ( ; . e . , when the intermittent heater shut off) to minimize error? resulting from the temperature fluctuation of the bath. n'hen the dilatometer was filled with pure solvent the level fluctuated about 0.1 mm. However, when readings were taken a t the same time during the heating cycle they agreed to within 0 . 2 rnm. Thus it appears t h a t the readings in the kinetic csperiments are good to about *0.2 mm. (27) R . Livingston in A. Weissberger's "Technique of Organic Chemistry," Vol V I I I , Interscience Publishers, Inc., New York, N. Y . . 1953, ChaDt. 2
KINETICSOF
July 5 , 1958
0rh.J-CLAISEN
I n the kinetic experiments the amount of substrate used was 1.5% by weight of the reaction mixture. This corresponds to a concentration of about 0.1 molar. Theallylaryl ethers were freshly distilled and the solvents were purified as follows. Diphenyl ether was purified according t o the method of Kincaid and Tarbell.10 N,N-Dimethylaniline was dried over fused potassium hydroxide prior to distillation and diphenylmethane was purified by fractionation. The other solvents (see Table V) were purified by standard methods. The specific first-order rate constants for the decrease in volume of the reaction mixtures were calculated by use of equation 1 in which Ro,R, and R.CO correspond to the readings a t zero time, time t and "infinity" time (;.e., after ten halflives). Since with dilute solutions the change in volume is a
k = (lit) In (Ro- I?-)/(& - R m ) (1) linear function of the percentage completion of the reaction,28 k is the first-order rate constant for the reaction. Usually about 25 readings were taken during a kinetic experiment and values of k were calculated for each of these readings. These values did not shov any trends over the range that the reactions were followed (to about 8 1% completion) and the average deviation was usually and a,,+
because the up+ constant has not been reported and it seems evident that the up value (based on one re2 the relationship between log k / k o and the Ham- action32)is much too low. The u p + constant for p mett up values is not linear. For those p-substitu- formyl calculated from the rate data in Table I V ents which have two u values, depending on the by use of equation 3 is 0.4G3. The data for the nature of the reaction, up values rather than up* two m-substituted allyl phenyl ethers included in values32were used in this plot. It is apparent that Table IV show positive deviations from this plot the up* values would show larger deviations than and have not been included. The deviation is esthe up values. I n this treatment the effect of the pecially serious for the m-methoxy ether. I n this substituents a t the m-position (where the allyl case the observed rate is about three times greater group becomes attached to the aromatic ring) has than that calculated from the urn+ value.ciG The been neglected. A substantially poorer correlation failure of the correlation for the nz-substituted (not shown) is obtained using urn constants instead ethers apparently is not due to the fact that the of up constants. substituent is ortho to one of the two possible posiAs illustrated by the lower plot in Fig. 2 the tions to which the allyl group can migrate. The data are correlated very well by the Hammett m-methoxy ether, which shows the largest deviarelationship (eq. 2 ) using Brown's up+ values.33 tion, is reported to give only the product resulting The equation for the least-squares fit is from rearrangement of the allyl group to the posilog k / k o = -0.510up+ + 0.078 (3) tion @ara t o the substituent,35 ;.e., 2-allyl-3The standard deviation from the least-square line methoxyphenol. Thus it appears that in this case (s) is 0.064 and the correlation coefficient ( r ) is the position ortho to the substituent is not involved 0.988.34 Rate data for allyl fi-formylphenyl ether in the reaction. As shown in Table VI there is little difference have not been included in this or the other plots between the u m and urn+ values. As indicated (31) L. P. Hammett "Physical Organic Chemistry," McGraw-Hill above, the correlation is very poor using values Book Co., Inc., iYew York, N. IT., 1910, Chapt. VII. and it is thus apparent that a m + would also give a (32) H. H. Jaffk. Chem. Ret's., SS, 1 U 1 (1953). (33) (a) H . C. Brown and Y. Okamoto, THISJOURNAL, 79, 1913 poor correlation. (1957); J . O r g . Chem., 22, 485 (1957); (b) N. C. Deno and W. L. Jaff6 has suggested32 that in cases (such as the Evans, THISJ O U R N A L , 79, 5801 (1957). present) where the substituents are both meta and (34) These quantities were calculated by the methods described log k / k " = pa,
in reference 32.
(2)
(35) F. Mauthner, J giakl Cheni , 102, 41 (1921)
HARLAN L. GOERIKGAND KOBERTR.JACOBSON
3% 4
VOl.
so
The data also were treated by equation 4 using constants instead of u constants. In this case pm:pr, = - 0 . i 5 and thus log k l k o was plotted against (up+ - 0 . T 5 n m t ) . As illustrated by the lower plot in Fig. 3 this relationship is indeed linear. This correlation is slightly better than that obtained using up+ constant; s = 0.046 and r = 0.994. The least-squares equation is u+
log k / k u = -0.678(upC -0.75 urn')
-0.049
In the latter correlation pp = -0.6'78 and pm = O.,jOS. This implies that the reaction is facilitated by electron supply to the P-position (the ether linkage) and electron removal from the m-position. However, this is clearly an illusion. This is shown by the fdct that the data are correlated by both equation 2 and equation 4 if ut constants are used. These two linear relationships are shown by equations 3 and 6. I
I
I
I
log k l k o = p u r , L
I
log k l k U =
+
~ p [ ~ p + ( ~ m / ~ i J ~ m + l
(3
(6)
Since these two equations hold, the relationship shown by equation 7 must also hold for the substituents used in the present work. d
0.6:
UI,+
5 0
ff
0.2-
-0.2-
-0.2 -1.4 Fig. 3.-Plots
-1.0
-
q+-o.7 5 G:
of log k/kO Z ' F Y S U S (ul, - a,,,) a n d (u,,+ 0.75,", &).
that obtained with the up values and is considerably poorer than that obtained with the up+ constants. The least-squares equation for this plot is log k / k u = - 1,C> I ( U ,-> u m ) - O,(J54
The standard deviation is 0 . 1 S and 7'
=
O.S!1S.
( 3 f i ) This ecliiation recently h : ~ heen . applied t n the iw!lrn-Claisen re-
arranxement o f p-substituted allyl p h e n y l ethers by 11.. S White. P! 0 1 , Abstracts uf the 131'nd Rleeting o f the American Chemical Suciety, September. 19.57. X e w York, N . Y . ,1). C C i i P. See also W. S . \Vhite. D . Gwynn, R. Schhtt, C . Gerard and 1%'. Fife, T H I SJ O U R N A L 80, 3iil (195X).
=
P'(U,,+
+ Cum+)
( 7)
In the latter equation p' and C are pp l p (equations 3 and 6) and pml'pp (equation G ) , respectively. Conversely, if the linear relationship shown by equation T holds it becomes apparent that any reaction series that can be correlated by the Hammett equation will also be correlated by the four parameter equation. Thus the significance of the two paratneters designated as pp and pm in the four I 6) is obscure. paratneter equation (and The relationship shown by equation 7 was tested using both u + and u constants. a'ith each type of constant the relationship was found to be linear. =\ plot of up+ versus up+ - O . T 5 a m + (C = -0.75) using constants for 23 substituents was linear, ,of = 1.28, Y = 0 . 9 i 2 , s = 0.15. .I plot of up aersus up - u,,,( C = - 1) using the values given in Table VI was also linear, p f = 1 . i 5 , r = 0.930, s = 0.16. This clearly shows that equation T holds and consequently it appears that equation 4 cannot be used for its intended purpose. The theoretical significance of the correlation of the rate data by the u + constants is not obvious. Perhaps the most important observation in this connection is that the data for the p-substituted ethers are correlated by u p + rather than urn+. The reaction is electron demanding, i.e., p is negative, and presumably the substituent effect is transmitted primarily through the 9- carbon-oxygen bond. The dependence of the rate on upc rather than um + implies that the transition state resembles the reactant more closely than the product. I t is not clear why in the present case the data are Correlated by u + constants instead of u constants. There are a n u ~ b e rof other reactions which are correlated better by U T than by u constants.33 However, for the most part these are cases where p is large and negative, i.e., the reaction is sensitive to substituent effects. The present reaction is relatively insensitive to substituent effects. The rather small effect of substituents is not sur-
July 5, 1958
3-385
CYCLIZATIONS OF 2-DEUTERIO-2’-CARBOXYBIPHENYLS
prising. If the reaction is considered t o involve simultaneous homolytic electron displacements p would be expected to be small. If simultaneous heterolytic displacements were involved the substituent effects a t the m- and +positions would be large but opposite in sign. Because of this cancellation the over-all effect would be small. 37 The present data do not rule out the possibility (37) Ambiguities concerning the nature, i.e., homolytic or heterolytic, and direction of electron displacements in concerted cyclic processes have been discussed by C. K. Ingold. “Structure and Mechanism in Organic Chemistry.” Cornell University Press, Ithaca, N. Y . , 1853, pp. 597, 637 and 641.
[CONTRIBUTION FROM
THE
that an intermediate may be involved in the rearrangement.38 However, there do not seem to be any reasons for suspecting t h a t this is the case. I t appears that the data are consistent with a concerted one-step cyclic mechanism. The geometry of the transition state for such a mechanism has been discussed p r e v i o ~ s l y . ~ ~ ~ ~ (38) If an intermediate is imolved the apparent first-order rate constant is a composite of three rate constants. The substituent and solvent effects can be predicted for the individual steps but not for the composite.
MADISON 6, WISCONSIN
SCHOOL OF CHEMISTRY, RUTGERS, THESTATEUNIVERSITY ]
Deuterium Isotope Effects in Some Acid-catalyzed Cyclizations of 2-Deuterio-2 ’-carboxybiphenyl BY DONALD B. DENNEYAND PETER P. KLEMCHUK’ RECEIVED DECEMBER 23, 1957
2-Deuterio-2’-carboxybiphenyl has been cyclized t o fluorenone with sulfuric acid, polyphosphoric acid and anhydrous hydrogen fluoride. The isotope effects for these reactions have been determined. Hydrogen fluoride cyclization exhibited a kH/kD of 3.02, whereas the other acid-catalyzed cyclizations had isotope effects in the range 1.13-1.46. The significance of these results is discussed with respect t o possible mechanisms for these substitutions.
Melander’s2pioneering studies on hydrogen isotope effects in aromatic nitration and bromination are now classics. Since his work was published many papers have appeared in which deuterium and tritium isotope effects have been measured during aromatic substitutions. These studies have led in general to a greater understanding of the mechanisms of these reactions and have specifically provided information concerning intermediates in the reaction sequence as well as to whether or not loss of hydrogen is involved in the rate-determining step. We wish t o report a t this time some measurements of deuterium isotope effects obtained during the acid-catalyzed cyclization of 2-deuterio-2‘carboxybiphenyl (I) to 4-deuteriofluorenone (Va) and fluorenone (Vb). As it can be seen, the isotope effects being measured here are those which have been termed intram~lecular.~This system D
0 \ 4 0 HI I
COOH D
0 Vb (1) Alfred P . Sloan Fellow in Chemistry, 1956-1957. (2) L. Melander, Arkiv Kemi, 2 , 213 (1950). (3) (a) W. hl. Lauer and W. E Noland, THISJOURNAL, 75, 3689 (1953); (b) T.G. Bonner. F. Bowyer and G. Williams, J. Chem. Soc., 2650 (1953); ( c ) P. B. D. de la Mare, T.M. Dunn and J T.Harvey, ibid., 923 (1957); (d) H. Zollinger, Helu. Chim. Acta, 8 8 , 1597 (1955); (e) U. Berglund-Larsson and L. Melander, Arkiv. K e m i , 6 , 21 (1953); ( f ) T. G. Bonner and J . M. Wilkins, J. Chem. SOL.,2358 (1955); ( g ) E. Grovenstein, Jr., and D. C. Kilby, THIS JOURNAL, 79, 2972 (1957). (4) K. B. Wiberg, Chem. Rens.. 66, 713 (1955).
was chosen because it provides a degree of flexibility not easily obtained when intermolecular isotope effects are measured. For example, it has been possible to measure isotope effects of very fast reactions by this technique, since the isotope effect is obtained simply by analyzing the mixture of Va and Vb for deuterium. We also have been able to measure isotope effects in heterogeneous reactions such as the Friedel-Crafts reaction by the use of this t e c h n i q ~ e . ~The isotope effects which we have observed can be found in Table I. I t is apparent TABLEI Catalyst
Time, min.
Teomp., C.
kdkn
Concd. sulfuric acid 96.63y0 sulfuric acid 86. FAY0 sulfuric acid Polyphosphoric acid Polyphosphoric acid Hydrogen fluoride
2-60 15 90 15 24 hr. 30
1 1 1 95 26 19
1.31+0.03 ‘1.31f . 0 3 1 . 1 3 + .02 1 . 3 1 i .03 1.16 f , 0 1 3.02 f .11
that all of the reactions studied exhibit some isotope effect. The range is considerable from a k H / k D of 3.02 in anhydrous hydrogen fluoride to 1.13 in 86.54% sulfuric acid. In order to accept these values as being meaningful, i t was necessary to prove that exchange was not taking place during the reaction. This sort of analysis has been accomplished for the sulfuric acid-catalyzed reaction. At various time intervals aliquots of the reaction mixture were removed and the fluorenone was isolated and analyzed for deuterium. All of the samples of fluorenone had the same deuterium content. If exchange was taking place with either the starting acid or the product, fluorenone, then there would be a steady decrease in the deuterium content of the fluorenone. Since this was not observed one knows that the isotope effects ob(5) Unpublished work.